This is a fork of the ORB-SLAM repository with some minor updates to publish camera pose on a seperate rostopic.

MORB-SLAM Monocular

This is a modified version of ORB-SLAM. The additional features contain:

Publishing of camera pose
Publishing of the current map points in the frame
Custom launch file for use in combination of our AutoPilot in Gazebo
Custom launch file for use in combination of our AutoPilot with real UAV's

Authors: Raul Mur-Artal, Juan D. Tardos, J. M. M. Montiel and Dorian Galvez-Lopez (DBoW2)

Current version: 1.0.1 (see Changelog.md)

ORB-SLAM is a versatile and accurate Monocular SLAM solution able to compute in real-time the camera trajectory and a sparse 3D reconstruction of the scene in a wide variety of environments, ranging from small hand-held sequences to a car driven around several city blocks. It is able to close large loops and perform global relocalisation in real-time and from wide baselines.

See our project webpage: http://webdiis.unizar.es/~raulmur/orbslam/

###Related Publications:

[1] Raúl Mur-Artal, J. M. M. Montiel and Juan D. Tardós. ORB-SLAM: A Versatile and Accurate Monocular SLAM System. IEEE Transactions on Robotics, vol. 31, no. 5, pp. 1147-1163, 2015. (2015 IEEE Transactions on Robotics Best Paper Award). PDF.

[2] Dorian Gálvez-López and Juan D. Tardós. Bags of Binary Words for Fast Place Recognition in Image Sequences. IEEE Transactions on Robotics, vol. 28, no. 5, pp. 1188-1197, 2012. PDF.

#1. License

ORB-SLAM is released under a GPLv3 license. For a list of all code/library dependencies (and associated licenses), please see Dependencies.md.

For a closed-source version of ORB-SLAM for commercial purposes, please contact the authors: orbslam (at) unizar (dot) es.

If you use ORB-SLAM in an academic work, please cite:

@article{murAcceptedTRO2015,
  title={{ORB-SLAM}: a Versatile and Accurate Monocular {SLAM} System},
  author={Mur-Artal, Ra\'ul, Montiel, J. M. M. and Tard\'os, Juan D.},
  journal={IEEE Transactions on Robotics},
  volume={31},
  number={5},
  pages={1147--1163},
  doi = {10.1109/TRO.2015.2463671},
  year={2015}
 }

#2. Prerequisites (dependencies)

##2.1 Boost

We use the Boost library to launch the different threads of our SLAM system.

sudo apt-get install libboost-all-dev

##2.2 ROS We use ROS to receive images from the camera or from a recorded sequence (rosbag), and for visualization (rviz, image_view). We have tested ORB-SLAM in Ubuntu 12.04 with ROS Fuerte, Groovy and Hydro; and in Ubuntu 14.04 with ROS Indigo. If you do not have already installed ROS in your computer, we recommend you to install the Full-Desktop version of ROS Fuerte (http://wiki.ros.org/fuerte/Installation/Ubuntu).

If you use ROS Indigo, remove the depency of opencv2 in the manifest.xml.

##2.3 OpenCV We use OpenCV to manipulate images and features. If you use a ROS version older than ROS Indigo, OpenCV is already included in the ROS distribution. In newer version of ROS, OpenCV is not included and you will need to install it. We tested OpenCV 2.4. Dowload and install instructions can be found at: http://opencv.org/

##2.4 g2o (included in Thirdparty) We use a modified version of g2o (see original at https://github.com/RainerKuemmerle/g2o) to perform optimizations. In order to compile g2o you will need to have installed Eigen3 (at least 3.1.0).

sudo apt-get install libeigen3-dev

##2.5 DBoW2 (included in Thirdparty) We make use of some components of the DBoW2 and DLib library (see original at https://github.com/dorian3d/DBoW2) for place recognition and feature matching. There are no additional dependencies to compile DBoW2.

#3. Installation

Make sure you have installed ROS and all library dependencies (boost, eigen3, opencv, blas, lapack).

Clone the repository:

 git clone https://github.com/PXL-IT/MPiF-ORB-SLAM.git ORB_SLAM

Add the path where you cloned ORB-SLAM to the ROS_PACKAGE_PATH environment variable. To do this, modify your .bashrc and add at the bottom the following line (replace PATH_TO_PARENT_OF_ORB_SLAM):
```
 export ROS_PACKAGE_PATH=${ROS_PACKAGE_PATH}:PATH_TO_PARENT_OF_ORB_SLAM
```
Build g2o. Go into Thirdparty/g2o/ and execute:
```
 mkdir build
 cd build
 cmake .. -DCMAKE_BUILD_TYPE=Release
 make 
```
Tip: To achieve the best performance in your computer, set your favorite compilation flags in line 61 and 62 of Thirdparty/g2o/CMakeLists.txt (by default -03 -march=native)
Build DBoW2. Go into Thirdparty/DBoW2/ and execute:
```
 mkdir build
 cd build
 cmake .. -DCMAKE_BUILD_TYPE=Release
 make  
```
Tip: Set your favorite compilation flags in line 4 and 5 of Thirdparty/DBoW2/CMakeLists.txt (by default -03 -march=native)
Build ORB_SLAM. In the ORB_SLAM root execute:

If you use ROS Indigo, remove the depency of opencv2 in the manifest.xml.
```
 mkdir build
 cd build
 cmake .. -DROS_BUILD_TYPE=Release
 make
```
Tip: Set your favorite compilation flags in line 12 and 13 of ./CMakeLists.txt (by default -03 -march=native)

#4. Usage

ORB_SLAM will receive the images from the topic /camera/image_raw. You can now play your rosbag or start your camera node. If you have a sequence with individual image files, you will need to generate a bag from them. We provide a tool to do that: https://github.com/raulmur/BagFromImages.

Tip: Use a roslaunch to launch ORB_SLAM, image_view and rviz from just one instruction. We provide an example:

Using Gazebo:

roslaunch orb_slam_gazebo.launch

Using an UAV:

roslaunch orb_slam_usbcam.launch

#6. The Settings File

ORB_SLAM reads the camera calibration and setting parameters from a YAML file. We provide an example in Data/Settings.yaml, where you will find all parameters and their description. We use the camera calibration model of OpenCV.

Please make sure you write and call your own settings file for your camera (copy the example file and modify the calibration)

#7. Failure Modes

You should expect to achieve good results in sequences similar to those in which we show results in our paper [1], in terms of camera movement and texture in the environment. In general our Monocular SLAM solution is expected to have a bad time in the following situations:

No translation at system initialization (or too much rotation).
Pure rotations in exploration.
Low texture environments.
Many (or big) moving objects, especially if they move slowly.

The system is able to initialize from planar and non-planar scenes. In the case of planar scenes, depending on the camera movement relative to the plane, it is possible that the system refuses to initialize, see the paper [1] for details.

PXLRoboticsLab/MPiF-ORB-SLAM

MORB-SLAM Monocular